This repository hosts GGUF-Imatrix quantizations for ResplendentAI/Sinerva_7B.
What does "Imatrix" mean?
It stands for Importance Matrix, a technique used to improve the quality of quantized models. The Imatrix is calculated based on calibration data, and it helps determine the importance of different model activations during the quantization process. The idea is to preserve the most important information during quantization, which can help reduce the loss of model performance, especially when the calibration data is diverse. [1] [2]
Steps:
Base⇢ GGUF(F16)⇢ Imatrix-Data(F16)⇢ GGUF(Imatrix-Quants)
Quants:
quantization_options = [
"Q4_K_M", "IQ4_XS", "Q5_K_M", "Q5_K_S", "Q6_K",
"Q8_0", "IQ3_M", "IQ3_S", "IQ3_XXS"
]
If you want anything that's not here or another model, feel free to request.
This is experimental.
For imatrix data generation, kalomaze's groups_merged.txt with added roleplay chats was used, you can find it here.
Original card information:
Decadent and rich in sensual prose, but beware, she is designed to humiliate and degrade her user when necessary.
Detailed results can be found here
| Metric | Value |
|---|---|
| Avg. | 70.38 |
| AI2 Reasoning Challenge (25-Shot) | 70.14 |
| HellaSwag (10-Shot) | 85.59 |
| MMLU (5-Shot) | 61.77 |
| TruthfulQA (0-shot) | 59.93 |
| Winogrande (5-shot) | 82.56 |
| GSM8k (5-shot) | 62.32 |